4 ms·
I'm guessing it means you can choose the precision (word length) of the computations, eg 8 bits, 16 bits, 32 bits, etc. This probably only applies to integer op
by p1esk 10y ago
I'm guessing it means you can choose the precision (word length) of the computations, eg 8 bits, 16 bits, 32 bits, etc. This probably only applies to integer ops, but if they could let us pack FP16 operands into 512 bit vectors, that would be awesome!
- rspeer 10y agoI hope they're going even farther and considering minifloats. 8-bit integers are easy but limiting and unexciting. 8-bit floats are exactly what some machine learning needs. If you're updating your weights enough, precision can just be wasted computation. A 512-bit-wide computation could be updating 64 very coarse-grained weights at a time.